The long term objectives of this program of research are to develop effective interventions for reducing problem drinking and associated problems among adolescents and young adults and to further enhance intervention approaches by identifying effective elements of treatment derived from cognitive behavioral social learning theory. Current approaches to behavior change in this area frequently rely on school-based primary preention programs that do not address cessation/reduction issues for adolescents who are already drinking, rarely address motivational issues related to use and abuse, and cannot target school dropouts. Recently, two studies have shown Motivational Interviewing (MI) to be effective with alcohol-involved adolescents when compared to a control or no intervention condition, but have shown greater harm-reduction effects than alcohol consumption effects. In addition, mechanisms of MI have not been elucidated. The major purposes of this study re to compare MI to a minimal contrast condition in which personalized feedback is provided, and to determine if additional booster sessions will enhance outcomes. The population is older adolescents who have been treated in an Emergency Department (ED) following an alcohol-related event. Thus, school dropouts, a high-risk population, will be included in the study. A 2 (MI versus Feedback Only) x 2 (two booster sessions versus no boosters) factorial design will be used to examine whether a MI combined with booster can effectively change subsequent alcohol use and alcohol problems. Experimental manipulations will be evaluated 6, 9, and 12 months after baseline intervention. The study design has several strengths: (1) it will enable an investigation of the main effects of MI versus Feedback Only, providing a more stringent test of the active ingredients of MI than our current competitive segment permitted; (2) it enables an evaluation of the effects of continued contact as a separate factor; and (3) it allows a test of the interaction between baseline intervention type and booster contact. A secondary purpose of the study is to use explicit mediational analyses, tested within a Structural Equation Modeling framework, to examine the hypothesis that stge of change, use of behavioral alcohol reduction strategies, and alcohol treatment seeking will mediate the relationship between intervention and outcome. Finally, the study will determine whether the diagnosis of alcohol abuse or dependence affects responsivity to our intervention. In addition to its potential contribut8ion to theory, the importance of this work is its potential for providing a cos-effective brief intervention at a "teachable moment" to increase high-risk patients' interest in reducing harmful drinking and related risk-taking behaviors.